EVT

• EVT
  • dataset
  • methods.peaks_over_threshold
  • methods.block_maxima
  • estimators.estimator_abc
  • estimators.hill
  • estimators.moment
  • estimators.gpdmle
  • estimators.gevmle
  • utils

Overview

Estimators and analysis for extreme value theory (EVT). The package is structured as follows. Example notebooks are provided as links.

• The Dataset object performs sanity-checks and contains plotting routines.

  • Generic plots of the dataset.

  • Maximum-to-sum plot.

  • Mean excess function.

• The peaks over threshold method.

  • Plot the tail.

  • QQ-plot against exponential.

  • Zipf-plot.

• The block maxima method.

  • Plot block maxima against the dataset.

• Estimators: calculate estimates and confidence intervals. Plotting routines for analysis.

  • Hill estimator. Plot against order statistics.

  • Moment estimator (Dekkers-Einmahl-De Haan). Plot against order statistics.

  • Maximum likelihood for the generalized Pareto distribution. Plot fit quality.

  • Maximum likelihood for the generalized extreme value distribution. Plot fit quality.

Documentation

Documentation is provided here. Example notebooks are provided here.

Installation

Releases are made available on PyPi. The recommended installation method is via pip:

>>> pip install evt

For a development setup, the requirements are in dev-requirements.txt. Subsequently, the repo can be locally pip-installed.

Indices and tables

• Index

• Module Index

• Search Page